Hryhorian, K.Volkov, K.Mazurok, Ihor Ye.Мазурок, Ігор ЄвгенійовичМазурок, Игорь Евгеньевич2020-05-122020-05-122020Інформатика, інформаційні системи та технології : тези доповід. 16-ї Всеукр. конф. студ. і молодих науковців, (Одеса, 24 квіт. 2020 р.). – Одеса, 2020.https://dspace.onu.edu.ua/handle/123456789/28030In the event of an emergency in public buildings, it is important for rescue services to know how many people were in the incident area. For example, the analysis of rubble and ruins in which victims may be located must be carried out with high speed and special precautions. When rescue operations are completed and there are no more people under the rubble or in the fire zone, other technical means can be used with less urgency and caution. This work is devoted to solving the problem of recognition and pairwise identification of incoming and outgoing people according to CCTV cameras at the inputs and outputs.enneural networkcomputer visionimage analysisMobileNetTransfer LearningArticle